The Bueda Semantic Analysis Engine claims to be able to translate user tags into categories, ontologies and structured content definitions. The system is based on research carried out by Carnegie Mellon University’s Language Technologies Institute.
Creating meaningful semantic data from unstructured sources (like user tags) is one of the fundamental problems for content repositories of all kinds – including Digital Asset Management systems so we were very interested to try it out. As with many solutions of this kind, as decision support (i.e. suggesting clean tags) they could be very useful and save a lot of time and improve cataloguing consistency, but they are still possibly a bit risky if fully automated.
When we tested the web based demo, the results were a bit hit and miss – but not at all bad and would probably be acceptable in a very large repository where a certain level of inaccuracy is more likely to be tolerated (certainly as compared with doing the work manually).
There is a demo on the Bueda site, API keys are available on request while the system is in private beta. The Semantic Focus site has access to a direct registration link also.
Related posts:
{ 2 comments… read them below or add one }
Thank you for the article. I am Bueda’s CEO and I would like to say that we are continuously improving our API and we are currently focusing on context disambiguation, which will improve the accuracy significantly. This should be deployed over the next few days as well as features like tag recommendation and an open source library for using the API. I look forward to get your feedback on the new features since feedback will shape our development effort to insure the usefulness of the service. One note is that the service infrastructure is really scalable and ready for large volume. Please don’t hesitate to contact me with any questions.
Vasco,
Thanks for your comment. Please keep up us updated on progress with Bueda, I’m sure many involved in metadata entry and cataloguing will be very interested in it.